Re: [PATCH RFC 13/31] tools/libxc: Wire a featureset through to cpuid policy logic

[Ian Campbell <ian.campbell@citrix.com>]

On Wed, 2015-12-16 at 21:24 +0000, Andrew Cooper wrote:
> Later changes will cause the cpuid generation logic to seed their
> information
> from a featureset.  This patch adds the infrastructure to specify a
> featureset, and will obtain the appropriate default from Xen if omitted.
> 
> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> ---
> CC: Ian Campbell <Ian.Campbell@citrix.com>
> CC: Ian Jackson <Ian.Jackson@eu.citrix.com>
> CC: Wei Liu <wei.liu2@citrix.com>
> ---
>  tools/libxc/include/xenctrl.h |  3 ++
>  tools/libxc/xc_cpuid_x86.c    | 90
> ++++++++++++++++++++++++++++++++++++++-----
>  2 files changed, 84 insertions(+), 9 deletions(-)
> 
> diff --git a/tools/libxc/include/xenctrl.h
> b/tools/libxc/include/xenctrl.h
> index 27e1f45..b44aec7 100644
> --- a/tools/libxc/include/xenctrl.h
> +++ b/tools/libxc/include/xenctrl.h
> @@ -2211,6 +2211,9 @@ int xc_cpuid_set(xc_interface *xch,
>                   char **config_transformed);
>  int xc_cpuid_apply_policy(xc_interface *xch,
>                            domid_t domid);
> +int xc_cpuid_apply_policy_with_featureset(xc_interface *xch, domid_t domid,
> +                                          uint32_t *featureset,
> +                                          unsigned int nr_features);

We should aim to eventually only have one apply policy interface here not
two, unless there is some good reason why xc_cpuid_apply_policy() should
remain?

AIUI passing featureset==NULL is going to be semantically identical to the
old interface, so I think we should just go with adding the new parameters
to the existing function.

> @@ -103,6 +109,32 @@ static int get_cpuid_domain_info(xc_interface *xch, domid_t domid,
>      info->hvm = di.hvm;
>      info->pvh = di.pvh;
>  
> +    /* Get featureset information. */
> +    rc = xc_get_featureset(xch, XEN_SYSCTL_featureset_host,
> +                           &host_nr_features, NULL);
> +    if ( rc )
> +        return rc;
> +
> +    if ( host_nr_features < XEN_FEATURESET_7c0 )
> +        return -EINVAL;

Could we assert this? I mean if Xen and libxc don't agree on this then
something went pretty wrong during the dev window.

> +
> +    info->featureset = calloc(host_nr_features, sizeof(*info->featureset));
> +    if ( !info->featureset )
> +        return -ENOMEM;
> +
> +    info->nr_features = host_nr_features;
> +
> +    if ( featureset )
> +    {
> +        memcpy(info->featureset, featureset,
> +               min(host_nr_features, nr_features) * sizeof(*info->featureset));
> +
> +        /* Check for trucated set bits. */

"truncated"

Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel